Why Diagnostic Analysis of your Data
Diagnostic analysis is an integral part of the model specification in economics. It helps to tell whether the concerned data has been well specified or not. The consequences of having a wrongly mis-specified model is extreme, as the results obtained would be wrong. Therefore, there is need to conduct the diagnostic tests for every model.
In addition to model specification, it is important to conduct the diagnostic tests, or what is referred to as the residual diagnostic tests. Some of the diagnostic tests offered at DataEdy include:
- Testing for Heteroskedaticity
- Testing for autocorrelation
- Testing for functional form
- Normality Test
- Correlogram and Q-statistics
- Testing for White Noise
The underlying concept behind any diagnostic analysis and tests is that if a model is correctly specified, then there are several weakly consistent estimators of the model parameters. Therefore, the estimates concerned should have a very little difference if the sample size is large. Most of the software provide the tools for conducting these diagnostic tests. However, the software used for analysis should also be applied in conducting the diagnostic tests.